Should AI generated images be sued for Copyright?

[SqueakyClean]

A case could be made it would be a net benefit, in that it met the demand but didn't cause harm to children.

My wife and I saw the response and briefly debated it last night. We kind-of came to the conclusion that it had the potential to go both ways. On one hand, yes, no child was harmed in the process, and if the weirdo who wants to see that can be satiated with AI generated material, it's a net positive. On the other hand, would the weirdo having access to this type of material easily embolden him to go after the real thing?

I suspect it's not an either/or situation. I'm guessing you would get a mixture of both.

As for the legality of it, I would guess on it depending on just how the images were generated. Can the "producer" of the content guarantee that all the images were created based upon fully clothed children and only naked adults that then the AI was able to blend together? Would that even matter?

 

[i'vegotwinners]

What Is the intended value to be protected by copyright?

If you ask AI to paint some Sun flowers, the product will always only be barely worth the paper it’s printed on. We aren’t talking Van Gough here.

"collectors/investors don't pay for "art" they pay for autographs".

the basic thought taken from a good movie i saw a couple decades ago about an art forger.

seems true to me to a great degree..

 

[larsIU]

My wife and I saw the response and briefly debated it last night. We kind-of came to the conclusion that it had the potential to go both ways. On one hand, yes, no child was harmed in the process, and if the weirdo who wants to see that can be satiated with AI generated material, it's a net positive. On the other hand, would the weirdo having access to this type of material easily embolden him to go after the real thing?

I suspect it's not an either/or situation. I'm guessing you would get a mixture of both.

As for the legality of it, I would guess on it depending on just how the images were generated. Can the "producer" of the content guarantee that all the images were created based upon fully clothed children and only naked adults that then the AI was able to blend together? Would that even matter?

Well said. Better than I would have anyway.

 

[SqueakyClean]

What Is the intended value to be protected by copyright?

If you ask AI to paint some Sun flowers, the product will always only be barely worth the paper it’s printed on. We aren’t talking Van Gough here.

There is also the marketing component. Like was shown in one of the posts above, it might easily get to the point where you can use AI generated images to sell handbags, or swimwear.

We could very soon be approaching the point where you can scan in a picture of a swimsuit that a client wants to market, and within a few hours of work, a visual effects artist can crank out an AI generated image of a generic supermodel languishing on the beach wearing that outfit. It will literally be tens of thousands of dollars in savings versus getting together the supermodel, the photographer, the support crew, transportation to the beach of all the personnel and the equipment, etc.

But if the AI generated image is an amalgamation of ten different supermodels, are they entitled to some of the NIL money?

 

[i'vegotwinners]

a basic table, if invented today, could probably produces a zillion patents.

you could patent every number of legs possible, sharp corners, rounded corners, everything in between, flat tops, every shape of legs, every length of legs, sharp edges, rounded edges, everything in between, every type material, every way to connect the components, etc.

can AI just start patenting/copyrighting/trademarking every concept on everything it can dream up..

all musical note combinations, or lyric combinations, all book plots, all possible movie plots, or any and all possible components therein?

all potential inventions not yet invented?

every possible mix of minerals or elements or chemicals?

then endless search everything created afterwards, human or AI based, for any and all infringements?

then in the same instant, initiate a suit against any infringement?

 

[larsIU]

a basic table, if invented today, could probably produces a zillion patents.

you could patent every number of legs possible, sharp corners, rounded corners, everything in between, flat tops, every shape of legs, every length of legs, sharp edges, rounded edges, everything in between, every type material, every way to connect the components, etc.

can AI just start patenting/copyrighting/trademarking every concept on everything it can dream up..

all musical note combinations, or lyric combinations, all book plots, all possible movie plots, or any and all possible components therein?

all potential inventions not yet invented?

every possible mix of minerals or elements or chemicals?

then endless search everything created afterwards, human or AI based, for any and all infringements?

then in the same instant, initiate a suit against any infringement?

I think @Noodle does something with intellectual property and patents? I could be wrong though.

I handle swimsuit review @ Target so not really my wheelhouse.

 

[Noodle]

a basic table, if invented today, could probably produces a zillion patents.

you could patent every number of legs possible, sharp corners, rounded corners, everything in between, flat tops, every shape of legs, every length of legs, sharp edges, rounded edges, everything in between, every type material, every way to connect the components, etc.

can AI just start patenting/copyrighting/trademarking every concept on everything it can dream up..

all musical note combinations, or lyric combinations, all book plots, all possible movie plots, or any and all possible components therein?

all potential inventions not yet invented?

every possible mix of minerals or elements or chemicals?

then endless search everything created afterwards, human or AI based, for any and all infringements?

then in the same instant, initiate a suit against any infringement?

Copyright and trademark protection like the above would be useless.

First, the verbs "copyrighting" and "trademarking" are oxymorons. You don't copyright something, as copyright exists the moment a work is created (you can register a copyright). Likewise, a trademark is a thing - a source identifier - that you use in selling goods or performing services. A trademark exists when the mark is used in commerce (and a trademark can be registered, but not before it has already become a trademark).

For copyrights, there is no infringement without proof of copying. There is no point in registering the copyright in "all musical note combinations, or lyric combinations, all book plots, all possible movie plots, or any and all possible components therein." If I write a song and no one ever hears it, it would be impossible for someone to infringe my copyright in that song. Even if I upload my song to YouTube, there can be no infringement if no one ever listens to the song.

For trademarks, no trademark rights exist without actual use of the mark in commerce - i.e., selling products or performing services under the mark. Having said that, there are hundreds, if not thousands, of seemingly random letter trademarks that have been registered in the U.S. recently by Chinese nationals. However, they've accomplished this by outright fabrication of images purportedly showing the random letter marks being used in commerce simply to obtain the registration. The Trademark Office has cracked down on this, but one net result is that there is a huge backlog of pending trademark applications. Up until about 2 years ago it took 2-3 months from filing before a trademark examining attorney (i.e., examiner) examined a new application. Currently, it's taking 10 months, and the delay is still slowly increasing.

Patents? Way too expensive for someone to do this. In the US, at least, one product already can generate numerous patents. But, there is a limit to how much someone is willing to spend. Outside of the US, it is unusual for one product to generate more than 2 patents in any given country. Plus, you can't simply patent "every possible mix of minerals or elements or chemicals." For starters, you have to establish that the thing you are seeking to patent has some real life use - not to mention that it is not an obvious variation of something already known.

 

[i'vegotwinners]

Copyright and trademark protection like the above would be useless.

First, the verbs "copyrighting" and "trademarking" are oxymorons. You don't copyright something, as copyright exists the moment a work is created (you can register a copyright). Likewise, a trademark is a thing - a source identifier - that you use in selling goods or performing services. A trademark exists when the mark is used in commerce (and a trademark can be registered, but not before it has already become a trademark).

For copyrights, there is no infringement without proof of copying. There is no point in registering the copyright in "all musical note combinations, or lyric combinations, all book plots, all possible movie plots, or any and all possible components therein." If I write a song and no one ever hears it, it would be impossible for someone to infringe my copyright in that song. Even if I upload my song to YouTube, there can be no infringement if no one ever listens to the song.

For trademarks, no trademark rights exist without actual use of the mark in commerce - i.e., selling products or performing services under the mark. Having said that, there are hundreds, if not thousands, of seemingly random letter trademarks that have been registered in the U.S. recently by Chinese nationals. However, they've accomplished this by outright fabrication of images purportedly showing the random letter marks being used in commerce simply to obtain the registration. The Trademark Office has cracked down on this, but one net result is that there is a huge backlog of pending trademark applications. Up until about 2 years ago it took 2-3 months from filing before a trademark examining attorney (i.e., examiner) examined a new application. Currently, it's taking 10 months, and the delay is still slowly increasing.

Patents? Way too expensive for someone to do this. In the US, at least, one product already can generate numerous patents. But, there is a limit to how much someone is willing to spend. Outside of the US, it is unusual for one product to generate more than 2 patents in any given country. Plus, you can't simply patent "every possible mix of minerals or elements or chemicals." For starters, you have to establish that the thing you are seeking to patent has some real life use - not to mention that it is not an obvious variation of something already known.

good info.

that said, could same AI not "publish" everything noted above on the internet, even on it's own site, for relatively little.

and offer a sale or download option for relatively almost nothing to cement "publishing".

that said, could AI doing this not cloud future claims, which i'd think would have a worth in of itself.

and the AI could handle all the legal work itself.

and do it's own analysis of the potential "worth" of any and all possibilities.

 

[Noodle]

good info.

that said, could same AI not "publish" everything noted above on the internet, even on it's own site, for relatively little.

and offer a sale or download option for relatively almost nothing to cement "publishing".

that said, could AI doing this not cloud future claims, which i'd think would have a worth in of itself.

and the AI could handle all the legal work itself.

and do it's own analysis of the potential "worth" of any and all possibilities.

Well, possibly. However, I think AI would have a hard time proving the accused infringer had access to the work they are accused of copying. Just saying it was on the Internet and therefore they had to have copied it generally would not be enough (unless, I suppose, alleged copy is so much like the original that there's a presumption of copying). In legal lingo, substantial similarity and showing defendant had access to the work gives a rebuttable presumption of copying. I don't think just saying it was on the Internet in some giant database would be sufficient to show access, but it's an interesting question.

 

[i'vegotwinners]

Well, possibly. However, I think AI would have a hard time proving the accused infringer had access to the work they are accused of copying. Just saying it was on the Internet and therefore they had to have copied it generally would not be enough (unless, I suppose, alleged copy is so much like the original that there's a presumption of copying). In legal lingo, substantial similarity and showing defendant had access to the work gives a rebuttable presumption of copying. I don't think just saying it was on the Internet in some giant database would be sufficient to show access, but it's an interesting question.

good info.

i obviously know nothing on the subject, just throwing stuff out there hypothesizing about AI.

long ago, my brothers and i used to jam together at Christmas every yr.

always wondered if posting recordings on YouTube offered any protection.

they were the talented ones, good song writers.

 

[CO. Hoosier]

There is also the marketing component. Like was shown in one of the posts above, it might easily get to the point where you can use AI generated images to sell handbags, or swimwear.

We could very soon be approaching the point where you can scan in a picture of a swimsuit that a client wants to market, and within a few hours of work, a visual effects artist can crank out an AI generated image of a generic supermodel languishing on the beach wearing that outfit. It will literally be tens of thousands of dollars in savings versus getting together the supermodel, the photographer, the support crew, transportation to the beach of all the personnel and the equipment, etc.

But if the AI generated image is an amalgamation of ten different supermodels, are they entitled to some of the NIL money?

Is it technically possible to create an AI gizmo that would produce more appealing and better quality “art” than any other AI gizmo? I don’t know. If so that giazmo might itself be able to have some kind of intellectual property protection.

 

[i'vegotwinners]

Is it technically possible to create an AI gizmo that would produce more appealing and better quality “art” than any other AI gizmo? I don’t know. If so that giazmo might itself be able to have some kind of intellectual property protection.

ask Hunter.

 

[SqueakyClean]

Is it technically possible to create an AI gizmo that would produce more appealing and better quality “art” than any other AI gizmo? I don’t know. If so that giazmo might itself be able to have some kind of intellectual property protection.

Ultimately, there is a user / technician that is creating the content. The "AI" at this point, is still just a really really advanced program that is taking input that the technician is giving it and spitting out a result. There is (currently....) not an AI that is deciding randomly to create the piece of art.

So the intellectual property is the technicians, not programs. Now, could one of these technicians make a piece of art, put it on the web, and then another technician uses it in his own AI generated art, and then the original technician tries to sue? I don't see why not.

 

[DANC]

You are correct. The 1970-71 team was the best ever. At one point IU swimmers held 9 of the 12 world records. Doc Counsilman was the draw. Heck, even when I graduated high school in 1981 he was still a draw, as two of my best friends growing up went to IU to swim even though at least one of them could have gotten a full ride at many other schools. At IU I think they swam in 3 Big 10 championship meets between the two of them.

50 years ago, IU swimming assembled the greatest team ever — in any sport

For a span of 19 days from late August to early September, out of 12 individual world records, Hoosier men held nine. “That will never be done again."

www.indystar.com

I used to go to the pathetic excuse for a workout area at the HPER building, where they had a few weights and an area along the wall where they had bars where you could do sit-ups.

This was the time - 1977/1978, I believe - when Councilman was training to swim the English Channel. I looked over at the old guy next to me, who had a "What's up, Doc?" t-shirt on.

It was The Man himself, doing sit-ups next to me. I had to actually start putting some effort into it - I didn't want to embarrass myself in front of him.

 

[Noodle]

I used to go to the pathetic excuse for a workout area at the HPER building, where they had a few weights and an area along the wall where they had bars where you could do sit-ups.

This was the time - 1977/1978, I believe - when Councilman was training to swim the English Channel. I looked over at the old guy next to me, who had a "What's up, Doc?" t-shirt on.

It was The Man himself, doing sit-ups next to me. I had to actually start putting some effort into it - I didn't want to embarrass myself in front of him.

LOL Doc was a very nice guy. Got to talk with him a few times at Big Tens, as he was good friends with our coach at Purdue (Doc Kahms). Hobie Billingsly (IU's diving coach back then) was one of the funniest people I have ever met. Not sure how they ever accomplished anything at diving practices with Hobie on deck.

Seems crazy that back then, at only 58, Doc was the oldest person ever to successfully swim the Channel. That no longer seems very old. The record for men is now 73 (women 71). Youngest is 11.

 

[larsIU]

Check out the Geoffrey Hinton interview on Zakaria (CNN). “Father of AI”

Thoughtful. He has concerns about whether we can control a thing which will be far more intelligent than humans ever could be. Can we create an ethical framework in which AI will exist? Or will human actors make that impossible?